How PlaylistPush took Google Ads ROAS from 8 to nearly 15 at the same spend

PlaylistPush connects independent artists with Spotify playlist curators and TikTok creators. Paid search is its main acquisition channel, and the account had grown in a way that wasted budget and blurred what was working.

• No account-level negative keywords: an estimated 30% of budget went to irrelevant clicks, and non-brand CTR sat well below benchmark.
• Performance Max was buying PlaylistPush's own brand searches at several times the cost of dedicated brand search.
• Conversion tracking did not attribute down to search terms, so Smart Bidding optimised on weak signals.
• In a category known for fake playlists, ads carried inconsistent claims about the number of artists served.

Plan: audit

A full Google Ads audit before changing anything, scored across tracking, wasted spend, structure, keywords, ads and settings. Beyond the wasted spend, it found:

• Performance Max buying PlaylistPush's own brand searches at four to six times the cost per acquisition of a dedicated brand campaign.
• Conversion tracking that stopped short of search-term level, so Smart Bidding was learning from weak signals.
• Non-brand click-through rate at about a third of benchmark, an average Quality Score below 6/10, and only 2 of 10 ad extensions in use.

Design: what to optimise for

A new account structure that separates brand, non-brand and Performance Max with shared negative lists, rolled out in stages with clear stop criteria. Each segment got its own targets, because brand and non-brand searches convert at very different rates:

• ROAS as the primary target, set separately for brand and non-brand, with a floor for the account as a whole.
• Cost per acquisition as a guardrail, capped below the break-even point.
• Conversion rate, cost per click and order value as the drivers behind ROAS, tracked separately so a market shift such as rising click prices is not mistaken for an account problem.
• Search-term relevance, Quality Score and ad strength as leading indicators.

Deploy: optimise

• Account-level negative keywords, and a brand exclusion on Performance Max so brand searches run through the cheaper brand campaign.
• Brand ads rewritten: over-pinning removed, two responsive search ads per ad group, and every claim checked against the live site.
• Non-brand campaigns relaunched with negative keyword sets and tested behind a spend gate. Tests that missed target were cut.

Adopt: report

• Weekly performance reports and monthly year-over-year reviews for the Head of Marketing.
• Every ROAS change broken down into order value, conversion rate and cost per click, so the team knows which lever moved.
• Alert-only Google Ads scripts that flag CPA or ROAS outside target, spend without conversions and budget pacing, plus a daily digest of account changes.
• One rule for ad copy: no number goes into an ad unless it is verified and dated.

• ROAS went from 8 to nearly 15 at the same spend level, an increase of about 85%.
• Conversion rate on paid traffic rose 25% year over year (July 2026), while cost per click rose 47%.
• Ad groups whose best ad was rated Poor went from 6 of 8 to none, and all 16 rewritten ads were approved.